Use of Spectral Biometrics for Aliveness Detection

Abstract

Numerous technologies are available for automatic verification of a person's identity. The authentication process usually involves verification of what a person knows (e.g., passwords, pass phrases, PINs), has (e.g., tokens, smart cards), is (e.g., fingerprint, hand geometry, facial features, retinal print, iris pattern), or generates (e.g., signature, voice). Use of something known by a person and use of something held by a person are two simple identification/verification solutions widely used today. Biometrics (also known as biometry) is defined as “the identification of an individual based on biological traits, such as fingerprints, iris patterns, and facial features” (McFedries, 2007), and relies on what a person is or can generate. Using something one knows requires only a good memory, but can on the other hand be easily overheard, seen, or even guessed. An item that one holds can be stolen and used or copied later. Using biometrics might at first seem to overcome these problems since fingerprints, iris patterns, etc. are part of one's body and thus not easily misplaced, stolen, forged, or shared. Indeed, biometrics technology is becoming a preferred standard for identification and authentication in ATMs, credit card transactions, electronic transactions, e-passports, airports, international borders, nuclear facilities and other highly restricted areas. Presently Europe leads the way but, the highest growth potential is forecasted to be in Asia as many Asian countries have already started adopting the technology. Its market size is estimated to be US$7.1 billion by 2012 (Bailey, 2008). Ironically however, this widespread acceptance of biometrics technology has been attracting the attention of attackers and has provoked interest in exploration of spoofing mechanisms against biometric systems. For example, the thousands of fingerprints that one leaves everywhere in one's daily life can be recovered and molded into artificial fingers for fooling biometrics devices based on fingerprint detection. In an experiment conducted by Matsumoto et al., eleven optical and silicon fingerprint sensors accepted artificial fingers in at least sixty percent of attempts (Matsumoto et al., 2002). Furthermore, with a commercially available high resolution digital camera, the iris pattern of a person's eye can be readily extracted from the person's facial picture and molded into contact lenses to be used to fool machines employing iris pattern recognition. An experiment conducted on two commercial iris recognition devices also showed that one of these devices could be fooled 50% of the time and the other 100% of the time (Matsumoto et al., 2002, 2004). Although susceptibility of most biometric system to spoofing have been experimented on fingerprint and iris recognition devices as these technologies are used in a variety of